On Wednesday 01 March 2006 23:19, Jim C. Nasby wrote:
> Probably the easiest way is to switch to using table partitioning and
> switch to using start_timestamp and end_timestamp, so that when you
> modify a row you update the old one setting end_timestamp to now() and
> insert the new row (all with
Probably the easiest way is to switch to using table partitioning and
switch to using start_timestamp and end_timestamp, so that when you
modify a row you update the old one setting end_timestamp to now() and
insert the new row (all within one transaction).
There are other ways to do it, but they'
Ricardo,
For future notice, there is another mailing list, PGSQL-PERFORMANCE,
devoted to questions like yours.
> select * from llamada where fecha='20030127' and tipo=1 and tiempo>0
> and gwdes like '64.7.127.14%' order by hora desc;
>
> It's very slow, it takes approximately 6 minutes to show m
=?iso-8859-1?B?UmljYXJkbyBKYXZpZXIgQXJhbmliYXIgTGXzbg==?= <[EMAIL PROTECTED]>
writes:
> I used PostgreSQL 7.0.3
Please consider upgrading to something more recent. 7.0 predates a lot
of work we've done on query optimization ...
regards, tom lane
---
